Characterization of a new Saccharomyces cerevisiae isolated from banana stems and its mutant with l-leucine accumulation for awamori brewing

ABSTRACT

We isolated a new strain of the yeast Saccharomyces cerevisiae, 35a14, from banana stems in Okinawa. This strain did not belong to any industrial yeast groups in a phylogenetic tree and produced high levels of alcohol. Furthermore, awamori, an Okinawa's traditional distilled alcoholic beverage, brewed with an l-leucine overproducing mutant derived from 35a14 showed a high concentration of isoamyl acetate.

Saccharomyces cerevisiae, banana yeast, comparative genomics, awamori, isoamyl acetate

Awamori, which is one type of shochu, a distilled alcoholic beverage made from steamed rice, is brewed primarily in Okinawa, Japan. With a history of over 600 years, this traditional distilled spirit is representative of Okinawan culture and industry. In awamori brewing, two microorganism species, the fungus Aspergillus luchuensis, as kuro-koji mold, and the yeast Saccharomyces cerevisiae, play essential roles for preparing the fermented mash (moromi) and producing ethanol, respectively, by multiple parallel fermentations, in which saccharification and fermentation take place simultaneously. After the completion of fermentation for about 2 weeks, the alcohol is mainly distilled under atmospheric pressure (Taira et al. 2012). The atmospheric distillation develops a rich and strong flavor. Thus, awamori has a unique aroma that is clearly distinguishable from the aroma profiles of other types of shochu (Fukuda et al. 2016). Since flavors of awamori are mainly affected by the metabolism of yeast cells, yeast breeding technology and natural isolation could contribute to the awamori industry. Previously, we isolated a mutant of the diploid awamori yeast strain 101-18, which accumulated a higher amount of l-leucine (Leu) selected from the l-leucine toxic analog 5,5,5-trifluoro-dl-Leu (TFL)-resistant colonies. The mutant strain 101-T55 has mutations in the LEU4 gene, which confers overproduction of a major flavor component, isoamyl acetate (Takagi et al. 2015). Recently, we isolated a novel strain of S. cerevisiae from hibiscus flowers in Okinawa, HC02-5-2, that produces high levels of alcohol and bred HC02-5-2 targeting the Leu synthetic pathway by conventional mutagenesis. The mutant strain T25 with Leu accumulation carries a mutation in the LEU4 gene and awamori brewed with T25 showed higher concentrations of isoamyl acetate than that brewed with the parent strain (Abe et al. 2019).

In this study, we aimed to obtain a new yeast strain that has the potential to improve the quality of awamori. Okinawa is the only place in Japan with a subtropical climate and has confirmed biodiversity, including animals and plants. Therefore, many microorganisms useful for industry are thought to inhabit the natural environment of Okinawa. Based on the above, we have attempted to isolate yeast strains from subtropical fruits and flowers in Okinawa such as strain HC02-5-2 isolated from the hibiscus flower (Abe et al. 2019). Since isoamyl acetate, which exhibits a banana-like flavor, is the preferred aroma component of awamori and contributes to its characteristics, the flavor of awamori is often associated with bananas. Among cultivated bananas in Okinawa, a small cultivar called “Shima-banana” is one of the popular fruits in local. We, therefore, focused on “Shima-banana” as a target for yeast isolation in the present study. Samples (76 in total) were collected from the leaves, fruits, and stems of “Shima-banana.” From these samples, strain 35a14 was isolated through enrichment culture and screening for ethanol production. The basic local alignment search tool search revealed that the 26S rDNA sequence of strain 35a14 was identical to that of the yeast S. cerevisiae. To investigate the potential of banana yeast 35a14 for awamori brewing, we performed a laboratory-scale awamori fermentation test (n = 3). First, the concentration of ethanol in moromi fermented with strain 35a14 reached 19.1%, whereas strain 101-18, a conventional awamori yeast, reached 19.2%. This indicates that strain 35a14 produces sufficient ethanol during fermentation. We then conducted next-generation sequencing for the genome of strain 35a14. As a result, 98.60% reads were mapped to the S288c reference genome, confirming that strain 35a14 belongs to S. cerevisiae. To examine the relationship between strain 35a14 and other yeast strains used for fermentation, a phylogenic analysis using whole genome information was performed by comparing single-nucleotide variants (SNVs) (Figure 1). Interestingly, the depicted phylogenic tree showed that strain 35a14 did not belong to any domesticated industrial yeast clades, such as sake, bread, or wine (Marsit et al. 2017), but was located near to the wild yeasts isolated from Malaysia, USA, and Japan. However, the evolutionary distances between 35a14 and wild yeasts were further apart than those between industrial yeasts. These results suggest that strain 35a14 is an endemic strain from Okinawa that evolved independently from other industrial yeasts and isolates.

Figure 1.

A phylogenic tree based on whole genome analysis of various S. cerevisiae strains. A single nucleotide variants-based phylogenic analysis was performed using the Neighbour-joining method. Based on information from the BioSample database (https://www.ncbi.nlm.nih.gov/biosample/), the isolation source, geographical information, and accession number are indicated for each strain. Scale bar represents 0.020 nucleotide changes per site.

Next, we measured the contents of aroma compounds in awamori produced in a laboratory-scale test according to the method described previously (Takagi et al. 2015; Abe et al. 2019). First, 4-vinyl guaiacol (4-VG) concentration observed in strain 35a14 (5.80 µg/mL) was ∼4 times higher than that observed in strain 101-18 (1.41 µg/mL). 4-VG is a precursor of vanillin known as a key flavor of aged awamori, and its concentration was quantified using HPLC system (Shimadzu) according to the method described previously (Abe et al.2019). Based on a previous report (Mukai et al. 2014), the variants in the PAD1 and FDC1 genes essential for the decarboxylation of phenylacrylic acids were compared. We confirmed that the intact and protein-coding PAD1 and FDC1 genes were present in the 35a14 genome, whereas a nonsense mutant was found in the FDC1 gene in strain 101-18. On the other hand, awamori brewed with strain 35a14 contained only about 60% of isoamyl acetate that brewed with strain 101-18 (Figure 2a). Similarly, the concentration of isoamyl alcohol, a precursor of isoamyl acetate, in strain 35a14 was 20% lower than that in strain 101-18 (Figure 2b). Isoamyl acetate is an important fruity aroma ester in the initial volatiles of awamori, and these compounds were analyzed using the electric nose of the Hercules II system (Alpha MOS) according to the method described previously (Abe et al.2019). Our results demonstrated that awamori brewed with strain 35a14 contained more 4-VG but less banana-flavored isoamyl acetate than that brewed with strain 101-18, respectively. Since strain 35a14 was isolated from the banana stem, it would be valuable if this strain could produce a large amount of isoamyl acetate. Therefore, we next bred mutants using strain 35a14 as a parent to improve the productivity of isoamyl acetate.

Figure 2.

Characteristics of strains 35a14 and BNNL80. (a) Isoamyl acetate and (b) isoamyl alcohol contents in moromi. Awamori brewing was conducted using strains 101-18, 35a14, and BNNL80. The white, grey, and black bars indicate moromi brewed with strains 101-18, 35a14, and BNNL80, respectively. (c) Intracellular Leu content in strains 35a14 and BNNL80. The Leu contents are represented as µmol per g dry cell weight (DCW). The values are the means and standard deviations of results from three independent experiments.

Since isoamyl acetate is synthesized from Leu via isoamyl alcohol in yeast cells, previous studies demonstrated that yeast strain with intracellular accumulation of Leu contributed to an increase in isoamyl acetate in alcoholic beverages such as sake (Ashida et al. 1987; Oba et al. 2005) and awamori (Takagi et al. 2015; Abe et al. 2019). We thus first isolated mutants resistant to TFL from the parent strain 35a14 by conventional mutagenesis to overcome the lower production of Leu-derived volatiles in awamori brewed with strain 35a14 compared with strain 101-18 (Figure 2a and b). It is known that overproduction of Leu can confer the TFL resistance to yeast cells (Baichwal et al. 1983), the TFL-resistant mutants are expected to synthesize a large amount of Leu in the cell. When strain 35a14 was randomly mutagenized by ultraviolet irradiation for 6 min (survival rate of ∼1.7%) and then plated onto the TFL (50 mg/L)-containing synthetic dextrose (SD) medium, ∼100 colonies were obtained as TFL-resistant mutants. Among these mutants, strain BNNL80 (24 µmol/g dry cell weight) accumulated 3.3-fold higher intracellular Leu than that of the parent strain 35a14 (7.0 µmol/g dry cell weight) when they were cultivated in liquid SD medium (Figure 2c). With an increase in the intracellular Leu, the concentrations of isoamyl acetate and isoamyl alcohol in moromi brewed with strain BNNL80 (16.3 µg/mL and 784.7 µg/mL) were 2.7- and 2.9-fold higher than those in the parent 35a14 (5.7 µg/mL and 291.9 µg/mL), respectively (Figure 2a and b). These Leu-derived flavor contents in moromi brewed with strain BNNL80 was 1.5- to 2-fold higher than those in moromi brewed with the industrial awamori yeast 101-18, indicating that the mutant strain BNNL80 has potential to be applied to brewing awamori with a rich banana flavor.

In the Leu biosynthesis, α-isopropylmalate synthase (IPMS) catalyzes the condensation of α-ketoisovalerate and acetyl-coenzyme A to form α-isopropylmalate. This condensation step is the rate-limiting because IPMS is subjected to feedback inhibition by Leu. In S. cerevisiae, two IPMS isozymes are encoded by the LEU4 and LEU9 genes. Previous studies reported that the introduction of amino acid substitutions to regulatory domain of Leu4, such as Ser542Phe, Ala551Val (Takagi et al. 2015), Gly516Ser (Abe et al. 2019), and Asn515Asp, Ser520Pro, Ala551Asp (Takagi et al. 2022), desensitized to Leu feedback inhibition, resulting in overproduction of Leu (Figure 3a). In order to elucidate a mechanism underlying Leu overproduction in strain BNNL80, we next analyzed the nucleotide sequence of the LEU4 gene and identified a homozygous mutation of guanine at position 1732 to adenine, leading to amino acid substitution of aspartate at position 578 with asparagine. In the previous study of sake yeast mutant, substitution of aspartate at position 578 with tyrosine in Leu4 reduced sensitivity to feedback inhibition by Leu, contributing to the overproduction of Leu thereby an increase in isoamyl alcohol and isoamyl acetate in sake (Oba et al. 2005). We, therefore, hypothesized that Asp578Asn substitution of Leu4 in strain BNNL80 would desensitize to Leu feedback inhibition similar to aspartate-to-tyrosine substitution. To confirm this, we introduced the mutation corresponding to Asp578Asn substitution to the LEU4 gene in the previously constructed expression vector of pYC130_LEU4 (Takagi et al. 2015), resulting in pYC130_LEU4^Asp578Asn. Yeast transformants harboring pYC130 vector, pYC130_LEU4 and pYC130_LEU4^Asp578Asn were cultivated in SD medium containing glutamate as a sole nitrogen source and G418, an antibiotic for maintaining the plasmids in the transformants. Subsequently, intracellular Leu contents were determined using an amino acid analyzer with ion-exchange chromatography and post-column ninhydrin derivatization (JLC-500/V2, JEOL, Tokyo, Japan). As shown in Figure 3b, yeast cells expressing the Asp578Asn variant (1.6 µmol/g dry cell weight) accumulated 1.6- and 1.4-fold higher intracellular Leu than that in yeast cells harboring the empty vector (0.98 µmol/g dry cell weight) and expressing the wild-type Leu4 (1.1 µmol/g dry cell weight), respectively. The result suggests that Asp578Asn substitution of Leu4 decreased sensitivity to Leu feedback inhibition, leading to high-level production of Leu.

Figure 3.

Analysis of Asp578Asn variant Leu4. (a) Domain organization of Leu4. Black triangles indicate the location of previously reported amino acid substitutions that conferred to removal of feedback inhibition by Leu, whereas white one represents that of Asp578Asn substitution identified in this study. (b) Intracellular Leu content in yeast cells harboring pYC130 vector (EV, white), pYC130_LEU4 (WT, grey), and pYC130_LEU4^Asp578Asn (D578N, black). The Leu contents are represented as µmol per g DCW. The values are the means and standard deviations of results from three independent experiments. Structure models of Leu4 constructed using MtIPMS without Leu (c) and with Leu (d). Asp578, Glu557, and Asp581, Lys489*, Arg495*, Asn515* are shown as a stick model in magenta, white, and yellow, respectively. Black dotted lines indicate putative hydrogen bonds. Residues between Glu557 and Asp581 are represented as a white ribbon, whereas those between Lys489* and Asn515* are yellow. The inhibitor Leu is shown as cyan stick model in panel (d).

In the crystal structure of IPMS from Mycobacterium tuberculosis bound with inhibitor Leu (PDB ID: 3FIG), two Leu binding sites are consisted of the regulatory domains from neighboring monomers (Koon et al. 2004). The inhibitor Leu is recognized by Asn532*, Leu535*, Ala565, Pro625, and Ile627, corresponding to Asn515, Ile518, Thr549, Glu577, and Val579 in Leu4 (asterisks donate residues in the neighboring monomer). Additionally, comparison of the crystal structure bound with Leu to that not including Leu (PDP ID: 3HPZ) indicates that the spatial locations of Asn532* and Ala565 are altered to recognize the inhibitor Leu. In the homodimeric structure model of Leu4 (Swiss-model depository P06208) that is constructed from MtIPMS without Leu (PDP ID: 3HPZ) as a template, Asp578 is located in the monomer-monomer interface of two neighboring monomers and the hydroxyl group in the side chain of Asp578 is expected to form intra-monomer interaction with Asp581 and inter-monomer interaction with Lys489* and Arg495*, respectively (Figure 3c). In contrast, Leu4 structure model that is constructed using MtIPMS bound with Leu (3FIG) as a template suggests that the alteration in the direction of the side chain of Glu577, Lys489*, and Arg495* cause a movement of Asn515* to recognize the inhibitor Leu (Figure 3d). Furthermore, the interaction between Asp578 and Arg495* would be disrupted by Leu binding, suggesting that Asp578 contributes to Leu-mediated conformational change. We thus hypothesize that substitution of aspartate at position 578 with asparagine may disrupt such interactions, leading to alteration in the spatial locations of Glu577 in the same monomer and Asn515 in the neighboring monomer, thereby preventing the binding of the inhibitor Leu. Previously identified Asp578Tyr substitution (Oba et al. 2005) may affect a similar influence on the local structure of Leu4, but further biochemical and structural analysis should be performed to clarify the detailed role of Asp578 in the allosteric regulation of Leu4.

In conclusion, the new isolate 35a14 from banana stems was confirmed to be a S. cerevisiae strain not directly related to other industrial yeasts or isolates. Strains 35a14 and its mutant strain BNNL80 were found to have favorable characteristics for the development of both the initial sented fruity flavor (isoamyl acetate) and the sweet flavor associated with aging (4-VG). Our data supported the practical application of these isolated yeast strains. Furthermore, we can now explore the solid molecular basis of the fermentation characteristics of these strains. Since distilled spirits contain many aroma components and their balance determines the quality, it is important to control multiple aroma components. This combinatorial approach of yeast isolation from nature and its breeding can be applied to the quality variation of alcoholic beverages in the fermentation industry.

Data availability

The data underlying this article are available in the article.

Author contribution

M.T. and H.T. conceived the study and designed the experiments. M.T., S.I., H.A., K.T, and Y.T. performed the experiments and analyzed the data. M.T., S.I, and H.T. wrote the manuscript. All authors read and approved the final manuscript.

Funding

This work was in part supported by Research Grants for Industry-Academia-Industry-Government Collaboration in Agricultural Chemistry and Small and Medium-sized Enterprises to H.T.

Disclosure statement

No potential conflict of interest was reported by the authors.

Author notes

These authors are equally contributed to this work.